Oracle America, Inc. v. Google Inc.REPLYN.D. Cal.August 29, 20111 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 GOOGLE’S REPLY ISO MSJ ON ORACLE’S COPYRIGHT INFRINGEMENT CLAIM CASE NO. 3:10-cv-03561-WHA 574160.04 KEKER & VAN NEST LLP ROBERT A. VAN NEST - #84065 rvannest@kvn.com CHRISTA M. ANDERSON - #184325 canderson@kvn.com 633 Battery Street San Francisco, CA 94111-1809 Telephone: 415.391.5400 Facsimile: 415.397.7188 KING & SPALDING LLP DONALD F. ZIMMER, JR. - #112279 fzimmer@kslaw.com CHERYL A. SABNIS - #224323 csabnis@kslaw.com 101 Second St., Suite 2300 San Francisco, CA 94105 Telephone: 415.318.1200 Facsimile: 415.318.1300 KING & SPALDING LLP SCOTT T. WEINGAERTNER (Pro Hac Vice) sweingaertner@kslaw.com ROBERT F. PERRY rperry@kslaw.com BRUCE W. BABER (Pro Hac Vice) 1185 Avenue of the Americas New York, NY 10036 Telephone: 212.556.2100 Facsimile: 212.556.2222 GREENBERG TRAURIG, LLP IAN C. BALLON - #141819 ballon@gtlaw.com HEATHER MEEKER - #172148 meekerh@gtlaw.com 1900 University Avenue East Palo Alto, CA 94303 Telephone: 650.328.8500 Facsimile: 650.328.8508 Attorneys for Defendant GOOGLE INC. UNITED STATES DISTRICT COURT NORTHERN DISTRICT OF CALIFORNIA SAN FRANCISCO DIVISION ORACLE AMERICA, INC., Plaintiff, v. GOOGLE INC., Defendant. Case No. 3:10-cv-03561-WHA GOOGLE INC.’S REPLY IN SUPPORT OF MOTION FOR SUMMARY JUDGMENT ON COUNT VIII OF PLAINTIFF ORACLE AMERICA’S AMENDED COMPLAINT Judge: Hon. William Alsup Hearing: 2:00 p.m., September 15, 2011 Case3:10-cv-03561-WHA Document368 Filed08/29/11 Page1 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 i GOOGLE’S REPLY ISO MSJ ON ORACLE’S COPYRIGHT INFRINGEMENT CLAIM CASE NO. 3:10-cv-03561-WHA 574160.04 TABLE OF CONTENTS Page I. INTRODUCTION ...............................................................................................................1 II. ARGUMENT.......................................................................................................................1 A. The Court should ignore or strike Oracle’s improper objections. ...........................1 B. Oracle must prove “virtual identity” between the works, which it cannot do..................................................................................................................1 C. The Java language APIs are methods of operation..................................................2 1. Oracle repeatedly mischaracterizes the APIs in an attempt to show creative expression where there is none. ............................................2 2. Oracle’s examples of “creativity” in the design of APIs are examples of form following function, and thus militate against copyrightability. ...........................................................................................3 a. The selection of elements to include in a set of APIs is not copyrightable. ............................................................................4 b. The structure for a set of APIs is not copyrightable. .......................5 c. The names of the elements in the APIs are not copyrightable....................................................................................6 3. Oracle improperly relies on dated law and other inapposite authorities.....................................................................................................6 D. Any creative expression in the Java language APIs has merged into their underlying ideas, or is unprotected scenes a faire...........................................8 E. Google’s fair use case is stronger than that of Sony v. Connectix, in which the Ninth Circuit held that Connectix’s use was fair. ...................................9 F. Oracle has identified no inconsistency between Google’s position and Google’s treatment of its own APIs.......................................................................10 G. Oracle offers no evidence of substantial similarity between the documentation of the Android core libraries and the Java language API documentation........................................................................................................11 H. The alleged copying in portions of 12 files out of several thousands of files is de minimis and thus not actionable.............................................................12 1. The alleged copying must be considered against the entirety of the Asserted Works to determine whether it is de minimis........................12 2. The rangeCheck function is not qualitatively or quantitatively significant...................................................................................................13 Case3:10-cv-03561-WHA Document368 Filed08/29/11 Page2 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 ii GOOGLE’S REPLY ISO MSJ ON ORACLE’S COPYRIGHT INFRINGEMENT CLAIM CASE NO. 3:10-cv-03561-WHA 574160.04 TABLE OF CONTENTS (cont'd) Page 3. The eight allegedly decompiled files are not qualitatively or quantitatively significant............................................................................14 4. The allegedly copied comments in the remaining two files are not qualitatively or quantitatively significant. ...........................................14 5. Google’s implementation of 37 Java language API packages is irrelevant to the de minimis analysis..........................................................15 III. CONCLUSION..................................................................................................................15 Case3:10-cv-03561-WHA Document368 Filed08/29/11 Page3 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 iii GOOGLE’S REPLY ISO MSJ ON ORACLE’S COPYRIGHT INFRINGEMENT CLAIM CASE NO. 3:10-cv-03561-WHA 574160.04 TABLE OF AUTHORITIES Page(s) Federal Cases Allen v. Academic Games League of America, Inc. 89 F.3d 614 (9th Cir. 1996) .........................................................................................................7 Apple Computer, Inc. v. Formula, Int’l, Inc. 725 F.2d 521 (9th Cir. 1984) .......................................................................................................7 Apple Computer, Inc. v. Microsoft Corp. 35 F.3d 1435 (9th Cir. 1994) ...................................................................................................1, 8 ATC Distribution Group, Inc. v. Whatever It Takes Transmissions & Parts, Inc. 402 F.3d 700 (6th Cir. 2005) .......................................................................................................3 Autoskill Inc. v. National Educational Support Systems, Inc. 994 F.2d 1476 (10th Cir. 1993) ...................................................................................................8 Bean v. McDougal Littell 669 F. Supp. 2d 1031 (D. Ariz. 2008) .......................................................................................12 Brown Bag Software v. Symantec Corp. 960 F.2d 1465 (9th Cir. 1992) .................................................................................................7, 8 CMAX/Cleveland, Inc. v. UCR, Inc. 804 F. Supp. 337 (M.D. Ga. 1992) ..............................................................................................8 Engineering Dynamics, Inc. v. Structural Software, Inc. 26 F.3d 1335 (5th Cir. 1994) .......................................................................................................8 Feist Pubs., Inc. v. Rural Tele. Serv. Co. 499 U.S. 340 (1991).....................................................................................................1, 5, 14, 15 Gates Rubber Co. v. Bando Chem. Indus., Ltd. 9 F.3d 823 (10th Cir. 1993) .........................................................................................................3 Incredible Techs., Inc. v. Virtual Techs., Inc. 400 F.3d 1007 (7th Cir. 2005) ...................................................................................................12 Johnson Controls, Inc. v. Phoenix Control Sys., Inc. 886 F.2d 1173 (9th Cir. 1989) .....................................................................................................6 Johnson v. Gordon 409 F.3d 12 (1st Cir. 2005)..................................................................................................11, 15 Lotus v. Borland 49 F.3d 807 (1st Cir. 1995), aff’d by an evenly divided court, 516 U.S. 233 (1996)........................................................................................................... passim Mattel, Inc. v. MGA Entm’t, Inc. 616 F.3d 904 (9th Cir. 2010) .....................................................................................................15 Matthew Bender & Co., Inc. v. West Pub. Co. 158 F.3d 674 (2d Cir. 1998) ........................................................................................................4 Mitel, Inc. v. Iqtel, Inc. 124 F.3d 1366 (10th Cir. 1997) ...................................................................................................9 Newton v. Diamond 388 F.3d 1189 (9th Cir. 2004) .............................................................................................13, 14 Case3:10-cv-03561-WHA Document368 Filed08/29/11 Page4 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 iv GOOGLE’S REPLY ISO MSJ ON ORACLE’S COPYRIGHT INFRINGEMENT CLAIM CASE NO. 3:10-cv-03561-WHA 574160.04 TABLE OF AUTHORITIES (cont'd) Page(s) Northwestern Nat’l Ins. Co. v. Baltes 15 F.3d 660 (7th Cir. 1994) .......................................................................................................11 Sega Enters. Ltd. v. Accolade, Inc. 977 F.2d 1510 (9th Cir. 1993) ........................................................................................... passim Sony Computer Entm’t, Inc. v. Connectix Corp. 203 F.3d 596 (9th Cir. 2000) ...........................................................................................9, 10, 11 Whelan Associates v. Jaslow Dental Lab 797 F.2d 1222 (3d Cir. 1986) ......................................................................................................6 Federal Statutes 17 U.S.C. § 102(b) ................................................................................................................. passim 17 U.S.C. § 106................................................................................................................................5 Federal Rules CIV. L.R. 7-3(a)................................................................................................................................1 FED. R. CIV. P. 56(c) ......................................................................................................................14 FED. R. EVID. 201(b) ........................................................................................................................9 Federal Regulations 37 C.F.R. § 202.1 .............................................................................................................................6 37 C.F.R. § 202.3(b)(3)..................................................................................................................12 37 C.F.R. § 202.3(b)(4)(i)(A) ........................................................................................................12 Other Authorities H.R. Rep. No. 1476, 94th Cong., 2d Sess. 57 (1976), reprinted in 1976 U.S.C.C.A.N. 5659 ......................................................................................................................3 Michael A. Jacobs, Copyright & Compatibility, 30 JURIMETRICS J. 91, 102 (1989)...........................................................................................................................................4 PAUL GOLDSTEIN, GOLDSTEIN ON COPYRIGHT (3d ed. 2011).................................................2, 3, 7, 8 Case3:10-cv-03561-WHA Document368 Filed08/29/11 Page5 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 1 GOOGLE’S REPLY ISO MSJ ON ORACLE’S COPYRIGHT INFRINGEMENT CLAIM CASE NO. 3:10-cv-03561-WHA 574160.04 I. INTRODUCTION Oracle’s Opposition concedes that, with the exception of portions of 12 out of several thousand files, it has no evidence of any copying by Google of any Oracle source code. Aside from those 12 files, Oracle’s copyright claim is premised on “non-literal” infringement and the incorrect notion that the Java language APIs—separate and apart from their implementations, which Google did not copy—are copyrightable. The Java language APIs at issue are unprotectable methods of operation, as Sun (now Oracle) itself told Congress in 1994. And even if there was any creative expression in those APIs, that expression has necessarily merged into the ideas represented by the APIs or is unprotectable scenes a faire, or its use by Google is a fair use as a matter of law. Oracle’s argument is largely based on the assertions that designing good APIs is difficult; that designing a highly functional set of APIs requires careful choices; and that given the alleged complexity of its APIs, they must contain some creative expression. But the Supreme Court rejected this type of “sweat of the brow” approach to copyrightability over twenty years ago. Feist Pubs., Inc. v. Rural Tele. Serv. Co., 499 U.S. 340 (1991). An extensive effort can no more transform methods of operation into copyrightable subject matter than industrious collection can turn facts into creative expression. Sun’s CTO was correct. Functional interfaces are not copyrightable. The Court should grant Google’s motion for summary judgment on Oracle’s copyright claim. II. ARGUMENT A. The Court should ignore or strike Oracle’s improper objections. “Any evidentiary and procedural objections to the motion must be contained within the [opposition] brief or memorandum.” CIV. L.R. 7-3(a). Oracle filed 8 pages of objections in addition to a 25-page opposition. The Court should reject Oracle’s improper objections. B. Oracle must prove “virtual identity” between the works, which it cannot do. Where “the range of protectable and unauthorized expression is [at most] narrow, the appropriate standard for illicit copying is virtual identity.” Apple Computer, Inc. v. Microsoft Corp., 35 F.3d 1435, 1439 (9th Cir. 1994). The Java language APIs are not protected at all, and Case3:10-cv-03561-WHA Document368 Filed08/29/11 Page6 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 2 GOOGLE’S REPLY ISO MSJ ON ORACLE’S COPYRIGHT INFRINGEMENT CLAIM CASE NO. 3:10-cv-03561-WHA 574160.04 the remaining allegedly copied expression is de minimis. Judged against the “virtual identity” standard—or even the “substantially similar” standard—there is no copyright infringement. C. The Java language APIs are methods of operation. The Java language APIs are the means by which developers can access the functionality of the Java language libraries. They are, in a very literal sense, methods for operating portions of the only works that Oracle pleaded in its Amended Complaint, namely versions 1.4 and 5.0 of the Java 2 Standard Edition platform1 (the “Asserted Works”). The APIs define the precise statements that must used for developers to access the functionalities in the Java language libraries. The APIs are thus unprotectable methods of operating the Java language libraries. 17 U.S.C. § 102(b); see also PAUL GOLDSTEIN, GOLDSTEIN ON COPYRIGHT § 10.5.1 (3d ed. 2011) (courts have “categorically excluded copyright protection for interface specifications”).2 1. Oracle repeatedly mischaracterizes the APIs in an attempt to show creative expression where there is none. Oracle repeatedly mischaracterizes the APIs, implying that they contain creative expression. Oracle is wrong. According to Oracle, the APIs “tell” how to use the libraries. Opp. at 2:27. Oracle is wrong. The APIs do not “tell” how to use the libraries, they are the means by which one uses the libraries; the documentation3 for the APIs “tells” how to use the libraries. According to Oracle, the APIs are the “blueprint” for the libraries. Id. at 3:5-6. Oracle is wrong. Unlike a blueprint, which states in detail how to build a structure, the APIs merely define the functions that the libraries implement. The proper analogy would be to a summary building 1 In addition to the APIs at issue, the Asserted Works include significant additional elements (such as the Java platform virtual machine and development tools) that are not the subject of Oracle’s copyright claim. 2 Oracle also claims that Google’s implementation (i.e., the collection of Java language libraries) is a derivative work of Oracle’s APIs, see Opp. at 6:5-9, even though it concedes that except for portions of 12 files there is no evidence of literal code copying. This type of bootstrapping is unsupportable. Copying an idea, fact or method of operation cannot be made actionable by alleging that the non-infringing code implementing these unprotectable elements is a derivative work. If it were, the non-infringing competitors would not have prevailed in Apple v. Microsoft, Sega v. Accolade, Sony v. Connectix or Lotus v. Borland. 3 Oracle’s claim that Google’s documentation infringes is addressed separately. See Part II.G, infra. Case3:10-cv-03561-WHA Document368 Filed08/29/11 Page7 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 3 GOOGLE’S REPLY ISO MSJ ON ORACLE’S COPYRIGHT INFRINGEMENT CLAIM CASE NO. 3:10-cv-03561-WHA 574160.04 plan that identifies the rooms that are to be included and other functional requirements, but does not describe implementation details. As Oracle’s expert acknowledges, the APIs are an “abstraction.” Mitchell Decl., Ex. 2 ¶ 23. According to Oracle, the APIs “describe” the fields and methods in the library classes. Id. at 3:11. Oracle is wrong. The documentation might “describe” the fields and methods, but an API merely identifies the fields and methods that are included. In short, the APIs are not the libraries themselves, and they do not “describe” or “tell” how to operate the libraries. Instead, the APIs are the methods of operating the libraries. And, by definition, methods of operation are not creative expression within the meaning of the Copyright Act. See 17 U.S.C. § 102(b).4 2. Oracle’s examples of “creativity” in the design of APIs are examples of form following function, and thus militate against copyrightability. Oracle argues that the design of good APIs requires skill and creativity, and that as a result its APIs are copyrightable. But that is not the law: “Original and creative ideas . . . are not copyrightable . . . .” ATC Distribution Group, Inc. v. Whatever It Takes Transmissions & Parts, Inc., 402 F.3d 700, 707 (6th Cir. 2005) (citing 17 U.S.C. § 102(b)). Moreover, functional aspects of a work are not copyrightable. Sega Enters. Ltd. v. Accolade, Inc., 977 F.2d 1510, 1524 (9th Cir. 1993) (citing 17 U.S.C. § 102(b)). The functional and factual aspects of a work may be copied, “as may those expressive elements of the work that ‘must necessarily be used as incident to’ expression of the underlying ideas, functional concepts, or facts.” Id. (quoting Baker v. Selden, 101 U.S. 99, 104 (1879)). Consistent with these core copyright principles, a work that is “largely functional” receives only weak protection under the Copyright Act. 977 F.2d at 1527. The aspects of API design that Oracle points to in support of its claim of creativity all focus on enhancing the functionality of APIs.5 Oracle argues that well-designed APIs are “easier 4 See also H.R. Rep. No. 1476, 94th Cong., 2d Sess. 57 (1976), reprinted in 1976 U.S.C.C.A.N. 5659, 5670 (“the expression adopted by the programmer is the copyrightable element in a computer program, and . . . the actual processes or methods embodied in the program are not within the scope of the copyright law”). At least one court has suggested that system architecture is one of the most common places where uncopyrightable processes will be found. See Gates Rubber Co. v. Bando Chem. Indus., Ltd., 9 F.3d 823, 837 (10th Cir. 1993). 5 The testimony from Google witnesses is not to the contrary. Some engineers prefer Case3:10-cv-03561-WHA Document368 Filed08/29/11 Page8 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 4 GOOGLE’S REPLY ISO MSJ ON ORACLE’S COPYRIGHT INFRINGEMENT CLAIM CASE NO. 3:10-cv-03561-WHA 574160.04 to learn and use” (Opp. at 3:23-24), and “almost disappear from sight” (id. at 4:1, quoting Swoopes Decl., Ex. 21). Oracle argues, in effect, that form follows function in well-designed APIs. But designing for functionality is the very antithesis of creative expression.6 a. The selection of elements to include in a set of APIs is not copyrightable. Oracle argues that its selection of API elements is protected. Opp. at 12:23-25 (“If the designer includes too little, the developers will not have the tools and flexibility they prefer. If the designer includes too much, the APIs become overwhelming and difficult to use.”). These arguments are immaterial because they do not make the APIs protectable. Determining what features are “must have” features may well require creative insights into the needs of programmers. Determining the point at which the cost of increasing difficulty of use outweighs the benefits of increased options may require refined judgment calls. But the feature set that results from this process is not a creative expression. Instead, it is a designer’s best guess at the optimal feature set for a product. Whether that notion is called an idea, a system, or a method of operation, it is not protected by copyright. See 17 U.S.C. § 102(b); see also Matthew Bender & Co., Inc. v. West Pub. Co., 158 F.3d 674, 682 (2d Cir. 1998) (finding “[t]he creative spark is missing where . . . the author made obvious, garden-variety, or routine selections . . . .”). Oracle’s theory would lead to absurd results. If feature sets in software were copyrightable, then competitors could not include the same feature set in their software. Moreover, Oracle concedes that the Android class libraries do not include all of the Oracle API elements, and that Android includes many API elements that the Oracle APIs lack. See Opp. at 20:27-21:1. Thus, under Oracle’s theory, merely having a partially overlapping feature set designing APIs over implementing, but that does not mean that APIs are more creatively expressive than the underlying code. Reply Astrachan Decl., Ex. 4 ¶ 32. Instead, it means that some engineers prefer coming up with ideas, while others prefer the different challenges that come from writing code in different ways to implement an idea. Id. 6 Industry practice has long reflected that APIs are uncopyrightable. For example, while AT&T required that vendors using its Unix implementation take a license, it understood that the specifications for the UNIX interfaces were not protected by copyright. See Michael A. Jacobs, Copyright & Compatibility, 30 JURIMETRICS J. 91, 102 (1989). Similarly, vendors of IBM- compatible personal computers relied on the unprotectability of interfaces in developing noninfringing alternatives to IBM’s BIOS. See id. Both Sun and Oracle have repeatedly relied on this principle. See id. at 100; Astrachan Decl., Ex. 1 ¶¶ 62-86. Case3:10-cv-03561-WHA Document368 Filed08/29/11 Page9 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 5 GOOGLE’S REPLY ISO MSJ ON ORACLE’S COPYRIGHT INFRINGEMENT CLAIM CASE NO. 3:10-cv-03561-WHA 574160.04 would constitute infringement. Such a result would allow developers to misuse copyright to claim monopoly control over entire product areas. This is precisely what the idea-expression dichotomy, on which copyright law is premised, is designed to prevent. Because the Java language APIs and their elements are functional, Google (and others) could have freely implemented all of the Java language APIs. But those elements are functional (and thus unprotectable) regardless of whether Google implemented all of them, or only some. Google’s decision to implement only a subset of the APIs cannot transform those APIs into protectable expression. Similarly, the presence in Android of additional APIs that are not part of the Java platform cannot change the unprotectable nature of the Java language APIs, and does nothing to detract from Google’s non-infringement argument. The Borland spreadsheet, for example, included commands not present in the Lotus 1-2-3 spreadsheet. See Lotus v. Borland, 49 F.3d 807, 810 (1st Cir. 1995), aff’d by an evenly divided court, 516 U.S. 233 (1996).7 b. The structure for a set of APIs is not copyrightable. Oracle also argues that “designing the appropriate structure” for a set of APIs results in creative expression. Opp. at 12:17. Oracle claims that the structure “is critical to the programmer,” and that, depending on the structure chosen, an API can be “your organization’s biggest asset” or can “even turn your users toward your competitor.” Id. at 13:6-19. These arguments are irrelevant to copyrightability. Not everything of value or competitive advantage is protected by copyright. Copyright protection is limited to original and creative expression of the types eligible for protection under the Copyright Act. The structure chosen by the designers of Oracle’s APIs is functional, not creative expression.8 To make use of the APIs, developers must know, among other things, the package, 7 Oracle’s claims that Google “fragmented” the Java APIs and that copyright should be used to enforce its “compatibility” licensing goals, see, e.g., Opp. at 8:9-15, 21:9-16, are likewise irrelevant. The Copyright Act does not protect against fragmentation or permit only “compatible” uses of unprotected elements. See 17 U.S.C. §§ 102(b), 106. Instead, everyone is allowed “to build freely upon the ideas and information conveyed by a work.” Feist, 499 U.S. at 350. Google is entitled to use as much or little as it wishes of Oracle’s unprotectable methods of operation. 8 A phone book could be organized alphabetically by last or first name, or geographically by street. The fact that there are different functional ways that data can be organized does not make any of those functional choices copyrightable. See Feist, 499 U.S. at 362-63. Case3:10-cv-03561-WHA Document368 Filed08/29/11 Page10 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 6 GOOGLE’S REPLY ISO MSJ ON ORACLE’S COPYRIGHT INFRINGEMENT CLAIM CASE NO. 3:10-cv-03561-WHA 574160.04 class, and subclass to which the methods belong; the parameters the methods take; the order of the parameters; the fields the data structures use; and how those fields are arranged. All of those features are necessary elements of the methods by which developers access the functionality in the libraries. Astrachan Decl., Ex. 1 ¶¶ 129-33; Reply Astrachan Decl., Ex. 4 ¶ 26. c. The names of the elements in the APIs are not copyrightable. Oracle’s opposition ignores the fundamental copyright principle that the names of the APIs and their elements are not copyrightable: “Words and short phrases such as names, titles, and slogans” are not copyrightable. 37 C.F.R. § 202.1 (emphasis added). There is no prohibition against using unprotected material, and that tautology is no less true if one uses many unprotected elements rather than just a few. Oracle’s fallback position—that the selection, coordination, and arrangement of unprotected elements can itself be expressive—fares no better, because the selection and structure of the APIs are not expressive, as explained above. See also Lotus, 49 F.3d at 816 (even if a software developer makes “some expressive choices” in choosing and arranging the menu items in its user interface, “that expression is not copyrightable because it is part of [the program’s] ‘method of operation.’”). 3. Oracle improperly relies on dated law and other inapposite authorities. Oracle’s authorities do not support its position. The Ninth Circuit’s pre-Feist decision Johnson Controls, Inc. v. Phoenix Control Sys., Inc., 886 F.2d 1173 (9th Cir. 1989), did not address software interfaces. Moreover, the Johnson Controls court’s statement that non-literal elements may be copyrightable where they are expression rather than ideas, id. at 1175, is not relevant here, because Oracle’s APIs are methods of operation, and thus by definition are not expression. 17 U.S.C. § 102(b). In addition, the analysis in Johnson Controls parallels the Third Circuit’s pre-Feist approach in Whelan Associates v. Jaslow Dental Lab, 797 F.2d 1222 (3d Cir. 1986), which, in the Ninth Circuit’s words, has been “widely—and soundly—criticized as simplistic and overbroad.” Sega, 977 F.2d at 1525-26. Oracle, in effect, invites the Court to repeat the Whelan court’s error of defining the “idea” of a software program unduly narrowly and then wrongly assuming that everything else must be creative expression. The CONTU Report also supports Google’s position, not Oracle’s. While the CONTU Case3:10-cv-03561-WHA Document368 Filed08/29/11 Page11 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 7 GOOGLE’S REPLY ISO MSJ ON ORACLE’S COPYRIGHT INFRINGEMENT CLAIM CASE NO. 3:10-cv-03561-WHA 574160.04 Report correctly states that the “written rules” of a game can be copyrighted even though they describe the process by which the game was played, CONTU Report at 21, this only means that the prose one uses to explain the mechanics of game may be copyrightable, even though the game mechanics themselves are not. See GOLDSTEIN ON COPYRIGHT § 2.15.1.3.9 As the CONTU Report explained, “[t]hat the words of a program are used ultimately in the implementation of a process should in no way affect their copyrightability.” CONTU Report at 21 (emphasis added). But while the implementation of a process (such as the implementation of an API) is potentially copyrightable, the process itself (like the abstract rules of a game, or an API itself) is not. 17 U.S.C. § 102(b). Nor does Apple Computer, Inc. v. Formula, Int’l, Inc., 725 F.2d 521 (9th Cir. 1984), another pre-Feist case, help Oracle. In that case, the Ninth Circuit rejected Formula’s argument that operating systems, which the court noted were “programs that are designed to manage the computer system,” are unprotectable methods of operation. Id. at 523-25. Formula conceded that the code it used was substantially similar to Apple’s code, see id. at 522-23, so there was no issue whether Formula was entitled to write code, not substantially similar to Apple’s code, that independently implemented the interfaces to Apple’s operating system. Apple sought “to copyright only its particular set of instructions, not the underlying computer process.” Id. at 525. The issue in this case—not addressed in Apple v. Formula—is whether the interfaces for accessing the Java language libraries are protected.10 And, for the reasons expressed herein, those interfaces are unprotected methods of operation. Oracle also mistakenly claims that in Brown Bag Software v. Symantec Corp., 960 F.2d 1465 (9th Cir. 1992), the Ninth Circuit “suggested” that screens, menus and keystrokes might be copyrightable. What the Ninth Circuit actually observed is that screens, menus and keystrokes might be suitable components for analytic dissection in the “extrinsic” prong of the test for 9 But see Allen v. Academic Games League of America, Inc., 89 F.3d 614, 617-18 (9th Cir. 1996) (applying merger doctrine and denying protection even to expression in game manuals). 10 Indeed, Oracle’s expert acknowledges the distinction between the expression in implementing code and the ideas of the APIs themselves when he states that the source code that implements the APIs is the “written expression” of the API. Mitchell Decl., Ex. 2 ¶ 55. Case3:10-cv-03561-WHA Document368 Filed08/29/11 Page12 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 8 GOOGLE’S REPLY ISO MSJ ON ORACLE’S COPYRIGHT INFRINGEMENT CLAIM CASE NO. 3:10-cv-03561-WHA 574160.04 substantial similarity. Id. at 1477. Analytic dissection is used to determine whether substantially similar elements are the result of protected or unprotected expression. Id. at 1475-76. The court’s reference to screens, menus and keystrokes as suitable components for analytic dissection does not suggest that those components are protectable, as Apple v. Microsoft made clear. Consistent with Apple v. Microsoft, the APIs and their selection and structure must be excluded from the infringement analysis because they are unprotectable. See supra, Part II.C.2.11 D. Any creative expression in the Java language APIs has merged into their underlying ideas, or is unprotected scenes a faire. To implement APIs that are compatible with the Java language APIs, Google had no choice but to replicate the precise details of the interfaces, just as GNU and Apache did with the GNU Classpath and Apache Harmony projects. See Mot. at 8:21-9:3.12 Nonetheless, citing Control Data Sys., Inc. v. Infoware, Inc., 903 F. Supp. 1316, 1323 (D. Minn. 1995), Oracle argues that because Sun had many options available when it designed the APIs, the particular choices Sun made do not merge into the ideas underlying the APIs. Opp. at 16:24-28. Courts in the Ninth Circuit, however, have rejected this approach. See GOLDSTEIN ON COPYRIGHT § 2.3.2.1 (observing that courts in the Ninth Circuit have held that “it is the range of expressive choice that existed at the time the competing product was created—not the range of expression that existed at the time the copyrighted work was created—that controls”). For example, in Sega, the Ninth Circuit noted that even specific computer code can be copied if it is the only means of accomplishing a task. 977 F.2d at 1524. At issue was code Sega had written and by which Sega-authorized games “unlocked” its Genesis gaming console. Id. at 1524 n.7. The court did not focus on whether the plaintiff had more than one option available to it when writing the “unlock” code, as it likely did. The court focused instead on whether the defendant at the time of the alleged infringement had alternatives to using the Sega code. Id. 11 Oracle also relies on Engineering Dynamics, Inc. v. Structural Software, Inc., 26 F.3d 1335, 1345-46 (5th Cir. 1994), which followed the district court reasoning in Lotus that was reversed by the First Circuit in 1995; and Autoskill Inc. v. National Educational Support Systems, Inc., 994 F.2d 1476 (10th Cir. 1993), and CMAX/Cleveland, Inc. v. UCR, Inc., 804 F. Supp. 337, 355 (M.D. Ga. 1992), both of which pre-date Lotus, Mitel, and Apple v. Microsoft. 12 Oracle does not dispute that it has not accused the GNU Classpath or Apache Harmony implementations of these same APIs of infringing its copyrights. Case3:10-cv-03561-WHA Document368 Filed08/29/11 Page13 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 9 GOOGLE’S REPLY ISO MSJ ON ORACLE’S COPYRIGHT INFRINGEMENT CLAIM CASE NO. 3:10-cv-03561-WHA 574160.04 Concluding that there were no such alternatives, the court rejected the argument that copying the unlock code infringed Sega’s copyright. Id. Oracle does not seriously dispute that to implement APIs that interoperate with the Java language APIs, Google had to adopt the structure and organization of those APIs. See Opp. at 17:1-3.13 Given the lack of choices available to Google, any arguable expression in the Java language APIs merges with the idea of providing APIs that are interoperable with the Java language APIs. See Lotus, 49 F.3d at 818 (“In the context of methods of operation” using unprotectable ideas “requires the use of the precise method of operation already employed”).14 E. Google’s fair use case is stronger than that of Sony v. Connectix, in which the Ninth Circuit held that Connectix’s use was fair. In Sony Computer Entm’t, Inc. v. Connectix Corp., 203 F.3d 596 (9th Cir. 2000), the Ninth Circuit held that the repeated intermediate copying of the entire code for Sony’s PlayStation BIOS was fair because the end product, Connectix’s Virtual Game Station (“VGS”), copied only unprotected elements of the BIOS. Among other things, VGS copied the “entry points” (i.e., the interface) into the BIOS and implemented 137 of the 242 functions supported by the BIOS. See Reply Kwun Decl., Ex. GG (Connectix’s Opening Appellate Brief) at 13, 18.15 Connectix deduced the required functionality by examining the parameters sent to the BIOS and the information returned by the BIOS. See id. at 13. In short, Connectix copied the APIs for over half of the functions in the Sony PlayStation BIOS. Moreover, during the development of VGS, Connectix repeatedly copied the entirety of the BIOS—i.e., Sony’s implementations as 13 Oracle suggests, without any supporting citation, that implementing the APIs at issue was not necessary to ensure compatibility with “Java.” Opp. at 17:18-19. But Oracle nowhere disputes that to support use of the Java programming language (which Sun long encouraged and over which Oracle claims no copyright) and provide interoperability with the well-known constructs in the Java language APIs, Google had to make use of the names, organization and structure of those APIs. 14 Oracle criticizes Lotus by noting that the Tenth Circuit took a different approach in Mitel, Inc. v. Iqtel, Inc., 124 F.3d 1366 (10th Cir. 1997). Opp. at 11:6-12. Mitel relied on scenes a faire, 124 F.3d at 1374-76, while Lotus relied on section 102(b), 49 F.3d at 815. However, under either approach, Oracle’s APIs are not protectable. 15 Google requests that the Court take judicial notice of Connectix’s Opening Appellate brief for the purpose of clarifying the facts and issues that were before the Ninth Circuit in that case. See FED. R. EVID. 201(b). Case3:10-cv-03561-WHA Document368 Filed08/29/11 Page14 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 10 GOOGLE’S REPLY ISO MSJ ON ORACLE’S COPYRIGHT INFRINGEMENT CLAIM CASE NO. 3:10-cv-03561-WHA 574160.04 well. See 203 F.3d at 605. Google, like Connectix, has developed an end product that includes implementations of a subset of the plaintiff’s APIs. Google, like Connectix, did so for the purpose of increasing interoperability. Oracle, like Sony, accuses the defendant of competing with its product. Here, however, Oracle has not accused Google of making unlawful intermediate copies of its implementation of the Java language APIs. And the undisputed facts show that Java has only become more popular since this lawsuit was filed. Compare Kwun Decl., Ex. A (Am. Compl.) ¶ 9 (alleging over 6.5 million Java developers) with Reply Kwun Decl., Ex. EE at 1 (recent Oracle press release touting over 9 million Java developers). Far from harming Oracle’s market, Android apparently “strapped another set of rockets to the [Java] community’s momentum,” just as Sun’s CEO predicted. Kwun Decl., Ex. L. The undisputed facts are therefore even more strongly in Google’s favor than they were in Connectix’s favor in Sony. The Ninth Circuit held, as a matter of law, that Connectix’s use of Sony’s BIOS APIs was fair. Sony, 203 F.3d at 608. A fortiori, Google’s use of the Java language APIs must also be fair. F. Oracle has identified no inconsistency between Google’s position and Google’s treatment of its own APIs. In a single sentence with no analysis, Oracle incorrectly claims that Google “asserts copyright . . . over its own APIs.” Opp. at 14:15-16. None of the three cited exhibits supports this conclusion. For example, the AdSense agreement states only that to the extent that there are intellectual property rights (including copyrights) in the “AdSense API Specifications,” those belong to Google. See Swoopes Decl., Ex. 23 at 6 (§ IV.10). Moreover, the agreement defines “AdSense API Specifications” to include the “Google-supplied implementations . . . of AdSense API,” not just the APIs themselves, which is what is at issue here. Id., Ex. 23 at 1 (§ I) (emphasis added); see also id., Ex. 24 at 4 (§ III.2) & Ex. 25 at 1. Oracle also suggests that Google seeks to prevent others from modifying the Android core libraries while reserving the right to implement libraries with functionality that differs from that of the Java platform libraries. See Opp. at 8:16-21. What Oracle ignores is that the Case3:10-cv-03561-WHA Document368 Filed08/29/11 Page15 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 11 GOOGLE’S REPLY ISO MSJ ON ORACLE’S COPYRIGHT INFRINGEMENT CLAIM CASE NO. 3:10-cv-03561-WHA 574160.04 consequence of modifying the Android core libraries is not a copyright lawsuit. Instead, devices that do not meet the Android compatibility guidelines simply “may not use the Android trademark.” Reply Kwun Decl., Ex. FF at 4. Regardless of whether the compatibility guidelines are met, “[a]nyone is welcome to use the Android source code . . . .” Id. G. Oracle offers no evidence of substantial similarity between the documentation of the Android core libraries and the Java language API documentation. The entirety of Oracle’s argument that the Android API documentation is infringing is its expert’s summary “conclusion” that Google’s and Oracle’s documentation are substantially similar, and a single cite to a side-by-side comparison of excerpts from that documentation. Opp. at 6:1-4. Its expert’s “analysis” again points to the same side-by-side comparison, and offers the following allegedly “similar” documentation: As an illustration, the Java API specification describes the method java.security.KeyPair.getPrivate as follows: “Returns a reference to the private key component of this key pair.” The Android spec uses the following prose: “Returns the private key.” Mitchell Decl., Ex. 1 ¶ 207. These two phrases are not substantially similar, and no reasonable jury could conclude otherwise. Indeed, this example only supports Google’s position: the only similarities between the two phrases are from the unprotectable facts about the java.security.KeyPair.getPrivate method that they are both describing. The Court is under no obligation to go beyond the single illustrative example that Oracle’s expert has offered. Cf. Northwestern Nat’l Ins. Co. v. Baltes, 15 F.3d 660, 662-63 (7th Cir. 1994) (“District judges are not archaeologists.”). However, if the Court were inclined to conduct its own comparison of the two respective sets of documentation without further guidance from Oracle, a brief review of the provided side-by-side comparisons demonstrates that any similarities arise only from unprotected facts about the APIs that are being documented. See generally Exs. E-F to Mitchell Decl., Ex. 1. The two sets of documentation are no more alike than two dictionaries that include and define the same words. In any event, the minimal expressive content in Oracle’s documentation is entitled to no copyright protection at all. “[C]opyright law protects original expressions of ideas but it does not safeguard either the ideas themselves or banal expressions of them.” Johnson v. Gordon, 409 Case3:10-cv-03561-WHA Document368 Filed08/29/11 Page16 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 12 GOOGLE’S REPLY ISO MSJ ON ORACLE’S COPYRIGHT INFRINGEMENT CLAIM CASE NO. 3:10-cv-03561-WHA 574160.04 F.3d 12, 19 (1st Cir. 2005) (emphasis added) (citing Feist, 499 U.S. at 345-51); Incredible Techs., Inc. v. Virtual Techs., Inc., 400 F.3d 1007, 1013 (7th Cir. 2005) (“utilitarian explanations” of a system “are not sufficiently original or creative to merit copyright protection,” or alternatively are protected “only against virtually identical copying”); see also Lotus, 49 F.3d at 815 n.9 (stating in dicta that a “strong argument” could be made that brief explanations of menu commands merge with the underlying idea of explaining the commands). H. The alleged copying in portions of 12 files out of several thousands of files is de minimis and thus not actionable. 1. The alleged copying must be considered against the entirety of the Asserted Works to determine whether it is de minimis. Oracle argues that the Court should evaluate infringement on a file-by-file basis, but the authority it cites requires exactly the opposite. Even if a registered work includes “copyrightable elements that are otherwise recognizable as self-contained works,” 37 C.F.R. § 202.3(b)(4)(i)(A) (emphasis added), where the claimant files a “single application” and pays a “single registration fee,” all of those copyrightable elements “shall be considered a single work.” Id. § 202.3(b)(4)(i) (emphasis added). Oracle did not register the files separately. In view of the registrations Oracle obtained and on which it bases its infringement claim, the allegedly infringed “works” are the Java SE 1.4 and 5.0 platforms, not the individual files. Moreover, the district court case on which Oracle relies, Bean v. McDougal Littell, 669 F. Supp. 2d 1031, 1034 (D. Ariz. 2008), discusses only the registration of a collective work. To claim copyright to “independent works in which copyright is being claimed and which appear within a larger work or within a collection of works,” the claimant must list the titles of all such independent works on a continuation sheet as part of the registration application. 37 C.F.R. § 202.3(b)(3). Oracle’s registrations do not list any individual files, see Kwun Decl., Ex. B, nor does Oracle provide any evidence that the individual files are anything but code fragments that are not recognizable as self-contained works. Oracle argues that comparing infringement only to the Asserted Works as a whole would lead to unjust results. This is incorrect for two reasons. First, Oracle, not Google, chose to register its works in the manner it did and to assert the specific copyright registrations that it has Case3:10-cv-03561-WHA Document368 Filed08/29/11 Page17 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 13 GOOGLE’S REPLY ISO MSJ ON ORACLE’S COPYRIGHT INFRINGEMENT CLAIM CASE NO. 3:10-cv-03561-WHA 574160.04 asserted. Second, Oracle is wrong to claim that this allows developers to “steal files at will from large software programs.” Opp. at 24:1-2. The de minimis defense applies only where the allegedly copied material is both quantitatively and qualitatively insignificant.16 The law requires that the alleged copying must be assessed in comparison to the Asserted Works as a whole. Newton v. Diamond, 388 F.3d 1189, 1195 (9th Cir. 2004). Applying that test, the limited alleged copying identified by Oracle is, as a matter of law, de minimis. 2. The rangeCheck function is not qualitatively or quantitatively significant. Oracle concedes that the rangeCheck function is quantitatively insignificant. Opp. at 24:15-17 (function is “only nine lines long” and is part of a file with “3,180 lines of code”). Oracle nonetheless claims it is qualitatively significant because it is called nine times in the Arrays.java file. This is a non-sequitur that fails to comprehend the meaning of “qualitative”; it is like saying that the letter ‘e’ is qualitatively significant in the English language because it is the most commonly used letter. See also Reply Astrachan Decl., Ex. 3 ¶ 56. The fact that a function is called repeatedly may have nothing to do with any qualitative significance, and everything to do with inefficient programming. Id. ¶ 57. And, in fact, that is the case here—due to limitations in the Java programming language, the author of Arrays.java was forced to call upon the rangeCheck code multiple times. Id. ¶ 58. “Thus the metric of number of calls is not a measure of the importance of rangeCheck, but rather of the inadequacies imposed by the Java language.” Id. Oracle’s only attempt to support its claim of qualitative significance is its expert’s speculation that there “may” be “more to it than meets the eye” to the nine lines of the rangeCheck function. Mitchell Decl., Ex. 2 ¶ 86. Oracle’s expert speculates that the order of the three error checks performed may be significant. Id. But he offers no explanation or evidence in support of this speculation, opining only that it is “possible” that “some amount of trial and error went into figuring out how to arrange the tests in this code so that the most informative error 16 If infringement were considered on a file-by-file basis—and, for the reasons stated, that would be error—Oracle’s damages and injunctive relief theories would need to be limited to the individual files identified by Oracle. Case3:10-cv-03561-WHA Document368 Filed08/29/11 Page18 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 14 GOOGLE’S REPLY ISO MSJ ON ORACLE’S COPYRIGHT INFRINGEMENT CLAIM CASE NO. 3:10-cv-03561-WHA 574160.04 condition is reported.” Id. Unfounded speculation cannot preclude summary judgment. See FED. R. CIV. P. 56(c). Finally, even were this speculation given credence—and given the lack of any evidentiary basis, it cannot be—it at best shows that any significance to the rangeCheck function derives from a functional choice arrived at through sweat of the brow. Whatever significance one might attach to the functional order of the three error checks performed, it has no qualitative significance under copyright law. See Sega, 977 F.2d at 1524; Feist, 499 U.S. 359-60. 3. The eight allegedly decompiled files are not qualitatively or quantitatively significant. Oracle effectively concedes that the eight allegedly decompiled files that Google received from a contractor are test files. Although Oracle asserts that the files were not part of the “test” directories in its works, Opp. at 25:15-16, that does nothing to illuminate whether they are in fact test files. Tellingly, Oracle’s expert does not and cannot dispute that they are test files. See Mitchell Decl., Ex. 2 ¶ 95. Oracle’s expert offers the circular speculation that because these files were part of Android, they must be important. Id. This speculation is belied by the fact that Google has removed these eight files from its Android distribution and did not replace them with anything else. Bornstein Decl. ¶ 8; Reply Bornstein Decl. ¶ 2. Oracle offers no evidence that the files are qualitatively significant. Its expert claims that the files are more than “dummy” files, Mitchell Decl., Ex. 2 ¶ 91, that the classes implemented have a “meaningful” function, id. ¶ 92, that some of the methods in these files include more program logic than others, id. ¶ 93, and that they implement certain Interfaces, id. ¶ 94. Even if these assertions were true, they offer no evidence that the eight files are qualitatively significant to the Asserted Works as a whole. Oracle’s expert’s analysis does nothing to bar summary judgment. Finally, Oracle offers no evidence to rebut Google’s showing that the eight files are quantitatively de mimimis, judged against the Asserted Works as a whole. 4. The allegedly copied comments in the remaining two files are not qualitatively or quantitatively significant. As Google has explained, the allegedly copied comments in the remaining two files are Case3:10-cv-03561-WHA Document368 Filed08/29/11 Page19 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 15 GOOGLE’S REPLY ISO MSJ ON ORACLE’S COPYRIGHT INFRINGEMENT CLAIM CASE NO. 3:10-cv-03561-WHA 574160.04 descriptive and functional. Mot. at 23:15; see also Astrachan Decl., Ex. 1 ¶¶ 172-73. Oracle does not dispute the mundane nature of the comments. See Opp. at 24:24-25:2. Oracle asserts only that the comments—which Google has since removed and not replaced, Bornstein Decl. ¶ 9; Reply Bornstein Decl. ¶ 3—are “significant” to the “works from which they have been taken.” Id. at 25:2. But the “works” Oracle refers to are the two individual files, based on Oracle’s incorrect argument about what “works” are at issue. See id. at 24:3-5. Oracle offers no evidence that the comments are qualitatively significant when measured against the Asserted Works as a whole. Indeed, it is questionable whether the comments include any protected expression in the first place, because “banal expressions” are not protected, even if original. Johnson, 409 F.3d at 19; see also Lotus, 49 F.3d at 815 n.9. And, again, Oracle offers no evidence to rebut Google’s showing that the comments in the two files are quantitatively de mimimis, judged against the Asserted Works as a whole. 5. Google’s implementation of 37 Java language API packages is irrelevant to the de minimis analysis. Oracle argues that the foregoing examples of, at most, de minimis copying must be considered together with Google’s decision to implement 37 Java language API packages. Oracle is again wrong. The infringement analysis proceeds only after the removal of unprotected elements via analytic dissection. Mattel, Inc. v. MGA Entm’t, Inc., 616 F.3d 904, 913 (9th Cir. 2010). As explained above, the APIs are not protected, and thus are not part of the infringement analysis at all. Because the alleged copying in the portions of the 12 files is, as a matter of law, quantitatively and qualitatively insignificant when compared to the Asserted Works as a whole, Oracle’s infringement claim fails. III. CONCLUSION For the foregoing reasons, Google respectfully requests that the Court enter summary judgment in Google’s favor on Count VIII of Oracle’s Amended Complaint. Dated: August 29, 2011 KEKER & VAN NEST LLP By: s/ Michael S. Kwun MICHAEL S. KWUN Attorneys for Defendant GOOGLE INC. Case3:10-cv-03561-WHA Document368 Filed08/29/11 Page20 of 20